PDA

View Full Version : Possible to animate tentacles procedurally in LW ?



blackmondy
08-17-2017, 11:46 AM
Anyone knows if Lightwave can do this without the tentacles intersecting one another ? Something like the Mimic creature in Edge of Tomorrow but of course not necessary as complex.

erikals
08-17-2017, 12:07 PM
IKBooster maybe...


https://www.youtube.com/watch?v=9-mPddEJMGo

edit; sorry realize i totally skipped reading important parts,
see if you can find ideas here > http://forums.newtek.com/showthread.php?140049-Tail-Whip

"procedurally" without intersection is going to be difficult i think.

squarewulf
08-17-2017, 12:31 PM
I gave a quick shot at it, could be tweaked for better results. I used dynamics with a point to point constraint with collision on and a procedural forcefield.

137699

Kind of rough, but i think it could work with enough patience.

erikals
08-17-2017, 01:11 PM
is it a lw 2015 scene file?

squarewulf
08-17-2017, 01:16 PM
is it a lw 2015 scene file?

yes

blackmondy
08-18-2017, 06:50 AM
I gave a quick shot at it, could be tweaked for better results. I used dynamics with a point to point constraint with collision on and a procedural forcefield.

Kind of rough, but i think it could work with enough patience.

Hi Squarewulf, appreciate your reply very much :)

I've checked out your attachment, very cool indeed. Been out of touch with Lightwave for quite a while myself, but would it be too much to ask if you can show the entire process via Youtube ?

Also, couple of questions :

1. Does the collision-detection work on the bones or on the geometry ?

2. Is it possible to have a null to change the direction of the tentacles in a non-uniform manner ? Like some sort of flocking, for lack of a better term.

3. Say I want to attached those tentacles to an object like a head (as a single object via welding), will these tentacles animation go haywire if attached to the head bone ?

Once again, thanks for everything.

- - - Updated - - -



IKBooster maybe...

edit; sorry realize i totally skipped reading important parts,
see if you can find ideas here >

"procedurally" without intersection is going to be difficult i think.

Thanks for replying.

The node-editor seems extremely powerful for that sort of thing somewhat, that's why I'm asking :)

squarewulf
08-18-2017, 10:33 AM
Hi Squarewulf, appreciate your reply very much :)

I've checked out your attachment, very cool indeed. Been out of touch with Lightwave for quite a while myself, but would it be too much to ask if you can show the entire process via Youtube ?


When i get back from lunch i'll see if I can refine this technique and maybe give you a quick video tutorial.

Collision detection is set on the bones themselves in the fx panel.

The only way I can think of to use a null in this manner would be to apply a forcefield or vortex to it. There may be a smarter way to have a null influence dynamics, but not that I know of.

If you attach the tentacles to something be sure to leave the object at 0,0,0 and move the root bone. From my experience, if your tentacles are separate objects from whatever you want it parented to, you will have to use the node editor in the motions panel to set the bones position to the parent. (item info, world position - object position)

This type of stuff isn't quite within my wheelhouse so i'm sure you could get better answers from other lightwavers here. But i'm curious how to do this as well so we can figure it out together.

MonroePoteet
08-18-2017, 05:20 PM
It's probably not what you're after, but I've attached a scene that just uses Deformation procedurals and the Bend and Taper Deformers. The base object was just a multi-segmented box using Magic Bevel to form the static, splayed tentacles. I created a Weight Map ranging from 0 at the base of the tentacles to 100% at the end of them. The Procedural is a large-scale Turbulence with an Alpha mode Gradient on the Weight Map.

137715137717

The Bend and Taper deformers are used to expand or collapse the tentacle "tree", or bend it around. Since the original tentacle tree on the static object aren't intersecting, I don't *think* the additive deformers can cause them to intersect. I can't see any in the sample scene.

Note that the Bend deformer affects the model from the BendBase to the BendHandle, which move in X and Y (the axis of the Bend being defined as Z in the deformer). The Taper deformer affects the model from TaperBase to TaperHandle, but uses the X and Y scale.

Also note that the position of the Bend and Taper deformers is NOT relative to the model they're affecting. They are based around (0,0,0) which is then translated to the object's pivot point.

mTp

P.S. As an aside, on the Mimicks in Edge of Tomorrow, I don't think I'd be able to see collision-detection failures. Those things move FAST!

sami
08-18-2017, 11:08 PM
It would be far easier to do this in Unity and export the output and then use it in LW.

blackmondy
08-19-2017, 07:56 AM
It's probably not what you're after, but I've attached a scene that just uses Deformation procedurals and the Bend and Taper Deformers. The base object was just a multi-segmented box using Magic Bevel to form the static, splayed tentacles. I created a Weight Map ranging from 0 at the base of the tentacles to 100% at the end of them. The Procedural is a large-scale Turbulence with an Alpha mode Gradient on the Weight Map.


The Bend and Taper deformers are used to expand or collapse the tentacle "tree", or bend it around. Since the original tentacle tree on the static object aren't intersecting, I don't *think* the additive deformers can cause them to intersect. I can't see any in the sample scene.

Note that the Bend deformer affects the model from the BendBase to the BendHandle, which move in X and Y (the axis of the Bend being defined as Z in the deformer). The Taper deformer affects the model from TaperBase to TaperHandle, but uses the X and Y scale.

Also note that the position of the Bend and Taper deformers is NOT relative to the model they're affecting. They are based around (0,0,0) which is then translated to the object's pivot point.

mTp

P.S. As an aside, on the Mimicks in Edge of Tomorrow, I don't think I'd be able to see collision-detection failures. Those things move FAST!

Thanks MonroePoteet, that was a very cool clip and appreciate your helpful reply.

I just feel that Lightwave has still a lot of untapped power under the hood, especially with the node editor. Tricky part is mastering it and knowing the combination of various plugins for some seriously cool effects similar to those seen in Houdini.

- - - Updated - - -


It would be far easier to do this in Unity and export the output and then use it in LW.

Could you explain that with a video ? :)

blackmondy
08-19-2017, 08:07 AM
Here's a link to the Mimic creature VFX. I'm intrigued by the growing tentacles part and that those tentacles are grouped together and moving as a larger tentacle. I think Lightwave could possibly do something like that, probably through some form of Lscript or a plugin like Relativity.

What you guys think ?


https://www.youtube.com/watch?v=iiKeTPL6HPk

erikals
08-19-2017, 09:00 AM
LightWave should be able to pull it off using endomorphs/bones/lattice/dynamics i think,
it's not an original concept though (already done elsewhere), so i see the challenge a bit redundant perhaps.

the technique itself though seems interesting,
can't help but think hardlink/metalink could be of help here, as it also can be use realtime...

where is Cageman??

erikals
08-19-2017, 09:27 AM
Lino + hardlink


https://www.youtube.com/watch?v=kWqRMimwizQ

i thought this could be done with metalink as well, but maybe not... (?)

seems possible, but realtime dynamics is trickier.
hmmm, coem to think, there is this, could be of help > https://www.youtube.com/watch?v=BpFzPBx6Ewk

prometheus
08-19-2017, 03:40 PM
I was also thinking of ik booster initially, I think with some of ryan roys tools and tutorials you may get somewhere, but ...I just did an initial test with bullet bone dynamics, but instead of going for point to point when having selecting the bones in order and applying type of bullet property, I went for hinge, since that has motors, so while all hinges are selected ..enable the motors, note, you would have to crank target motor and motor force up to thousands for it to kick in properly..and it should start twitching and curling, you can then either use noisy channels in the envelope modifier or use textures for that motor, to further note, you should also switch to select all bones and increase collision margin in order for the tentacle to collide with itself or other tentacles, this way you do not need additional forces to procedurally move the tentacles, and it would also not be the same as a motor driven tentacle which behaves more like a muscle tentacle that lives by itself rather than swaying in the wind based on some other force.
Note..the high value of motor force and target motor speed may be reduced if you lower the initial density values of the rigid body bones, they are by default set to 1000 ..so that is correlating.

I Have to try with more than one tentacle tomorrow.
But go ahead..make one tentacle and make it a hinges instead of point to point.

Mr Rid
08-19-2017, 06:21 PM
Awhile back, I tried it with IKB, bone dynamics and wind effectors. The problem is that bones do not self-collide. And Bullet lacks effectors.

Surrealist.
08-19-2017, 08:04 PM
If asked to do something like this I would simplify the approach to something more like bones and animating secondary motion.

My next approach would be IKB dynamics for the secondary motions.

In general I would only worry about intersections that were in the camera view. And plan a way to edit those by hand with the bones. (Motion Mixer layer?)

Or bake out final shots to mdd and edit in C.S.

bazsa73
08-20-2017, 12:59 AM
LW is totally unapt for this kind of animation. Maya or Houdini.

Surrealist.
08-20-2017, 01:19 AM
But with all due respect, while I personally would prefer to do something like this in a more adapt program. (I use Maya, Houdini is not so simple to Master, but I think you are right - and many other options as well). But lets keep this on topic. He asked about doing this in LightWave. While I agree there are other better tools. Why not simply stay focused on how to do something like this in LightWave?

MichaelT
08-20-2017, 01:28 AM
What you're looking for is a mixed solution:
https://www.youtube.com/watch?v=xjSio43UZFY
https://www.youtube.com/watch?v=Wnjye8CPouw
https://www.youtube.com/watch?v=5WNZO5uA4gs
etc..

blackmondy
08-20-2017, 03:29 AM
LW is totally unapt for this kind of animation. Maya or Houdini.

Have you tried it yourself ?

I believe Lightwave is capable of doing it, just that it will probably take a lot more effort and workarounds.

prometheus
08-20-2017, 06:22 AM
Awhile back, I tried it with IKB, bone dynamics and wind effectors. The problem is that bones do not self-collide. And Bullet lacks effectors.

I guess you would have to turn your tentacle mesh to a collision object, that could work roughly, but it would be problematic to get accurate collisions anyway and above else..it would be soo slow.
So my bet would be on bullet bone dynamics, at least its fast and you can set rigid bodies collision size for every bone.


What about bullet no effectors, what would you need that for with bullet? is there something that bullet forces or a bullet kinematic item canīt do?

prometheus
08-20-2017, 06:37 AM
Lorenzo Zitta made the octopuss with bullet bone dynamics, featured for the lw 2015 release, take a look at that..and maybe contact him for advice...
He used cone twist for the constraints, I tried that and it seems to be working ok.

0:10 in the clip...

https://www.youtube.com/watch?v=j9oq1m33bzU

blackmondy
08-20-2017, 08:01 AM
I have a LWO object attached in case anyone wants to try out his/her theory. :)

- - - Updated - - -


What you're looking for is a mixed solution:
https://www.youtube.com/watch?v=xjSio43UZFY
https://www.youtube.com/watch?v=Wnjye8CPouw
https://www.youtube.com/watch?v=5WNZO5uA4gs
etc..

Thanks for the heads-up... :)

Mr Rid
08-21-2017, 12:19 AM
Years ago, I pondered how to animate tentacles like the Davy Jones character in Pirates of the Caribbean. How do you control tentacles with keyframes, collision dynamics, and random procedural wiggliness, all simultaneously?
I did these tests using IK Boost, bone dynamics, and wind effectors. I wish bones would self-collide! But bone collisions are imperfect, and the tentacles pass thru each other and themselves.

I tried a similar setup with ClothFX (and weights), instead of IKB, but I could not get the mesh to not jitter or deform oddly.

Negative Explosion winds attract the tentacles. Vortex and Turbulence winds make the wriggle.


https://www.youtube.com/watch?v=zq7Q0frnIeM


Lorenzo Zitta made the octopuss with bullet bone dynamics

The limitation with Bullet is there are no effectors besides gravity. You see how the octopus tentacles only trail.

But I do not see a good way to do this in LW.

blackmondy
08-21-2017, 04:01 AM
Years ago, I pondered how to animate tentacles like the Davy Jones character in Pirates of the Caribbean. How do you control tentacles with keyframes, collision dynamics, and random procedural wiggliness, all simultaneously?
I did these tests using IK Boost, bone dynamics, and wind effectors. I wish bones would self-collide! But bone collisions are imperfect, and the tentacles pass thru each other and themselves.

I tried a similar setup with ClothFX (and weights), instead of IKB, but I could not get the mesh to not jitter or deform oddly.

Negative Explosion winds attract the tentacles. Vortex and Turbulence winds make the wriggle.

The limitation with Bullet is there are no effectors besides gravity. You see how the octopus tentacles only trail.

But I do not see a good way to do this in LW.

Thanks for the insightful reply... :)

erikals
08-21-2017, 04:43 AM
The limitation with Bullet is there are no effectors besides gravity.
haven't tried bullet much, maybe one could use procedural displacement instead of wind?

prometheus
08-21-2017, 10:15 AM
haven't tried bullet much, maybe one could use procedural displacement instead of wind?

Donīt see how that is a replacement for the lack of effectors?

prometheus
08-21-2017, 10:18 AM
Years ago, I pondered how to animate tentacles like the Davy Jones character in Pirates of the Caribbean. How do you control tentacles with keyframes, collision dynamics, and random procedural wiggliness, all simultaneously?
I did these tests using IK Boost, bone dynamics, and wind effectors. I wish bones would self-collide! But bone collisions are imperfect, and the tentacles pass thru each other and themselves.

I tried a similar setup with ClothFX (and weights), instead of IKB, but I could not get the mesh to not jitter or deform oddly.

Negative Explosion winds attract the tentacles. Vortex and Turbulence winds make the wriggle.


https://www.youtube.com/watch?v=zq7Q0frnIeM



The limitation with Bullet is there are no effectors besides gravity. You see how the octopus tentacles only trail.

But I do not see a good way to do this in LW.

I am not sure either, maybe kinematic objects? ideally if the layout spline control tools could work alongside bullet dynamics, but I am doubtful that will work satisfactory?
Now would be a good idea for Lino to take a coffe break and investigate it, or perhaps if Royan Roye could jump in and recommend something.

perhaps it is possible to add kinematic objects parented to certain bone parts in a chain, or change certain bones to be of a certain type that makes it more keyframe editable?

erikals
08-21-2017, 11:07 AM
Donīt see how that is a replacement for the lack of effectors?

seen here, displacement wind


https://www.youtube.com/watch?v=bliPxn5e7mM

Mr Rid
08-21-2017, 05:58 PM
seen here, displacement wind



I suggest Normal Displacement will give a more organic result.

jwiede
08-21-2017, 06:30 PM
seen here, displacement wind

The technique in that video works because all the leaves are in random absolute orientations (yet share common UVs & root-v-tip local orientation), so randomly (and relatively) displacing the leaves' end vertices (with reducing falloff towards branch end) results in them all moving in different absolute directions at tip end while remaining "still" at the branch end (due to the falloff).

Can you explain how you'd apply that technique to a tentacle's geometry to yield tentacle-like movement? I'm not seeing how that would work.

Just displacing the end with a falloff to the "root" of the tentacle doesn't seem like it'll give a convincing "tentacle-like" movement (for the same reason the leaves didn't curl and bend in the tutorial), nor would even adding random (relative) displacement of the tentacle vertices along the length seem to yield such a movement.

Specifically, please clarify how you'd apply that displacement technique, to which tentacle vertices, and what orientation of weight-map falloff you'd use tentacle tip-vs-root?

erikals
08-21-2017, 08:14 PM
it is rater a workaround, with limitations.

one would have to use global displacements, with the limitations that would cause.

it's limited, though. maybe too limited. just wondered if others had had the same or a related idea.

Surrealist.
08-22-2017, 03:30 AM
Looking at the mimic example again it seems like it could be done with traditional rigging and morphs along bones. I really don't see a need to do any if this effect with dynamics. At least not with LW limitations.

I would say making a custom rig and maybe some plugin develolment to make controlling the morphs easier. Just guessing. But from what I actually see of the effect showcased looks very doable in LW with a combination of effects and good old hand animation techniques.

In the effect shown there is nothing procedural it seems. I mean in terms of displacements or secondary motion. The procedural aspect is in the generation of tenticles along a path. And that can be done with morph along bones.

RebelHill
08-22-2017, 09:09 AM
Yep, you can pretty much forget doing this well in LW with the standard native tools, you can try any number of combinations that'll only involve a lot of work with less than great end results... too many complexities involved to solve that way.

Of course, it's totally doable, but would require you to write your own custom motion and/or displacement plugins to handle the specifics of the problem in the way desired.

jeric_synergy
08-22-2017, 09:41 AM
How were the originals animated?

blackmondy
08-22-2017, 09:43 AM
OK here's the details on the Mimics creature effects :

https://www.fxguide.com/featured/live-die-repeat-the-effects/


What's the scoop about Fabric Engine ? Would be nice for Lightwave to have it's own "SoftImage ICE" for doing this kinda complex stuff. Any idea if the two work together ?

RebelHill
08-22-2017, 10:17 AM
And there you go...

“We didn’t want animators spending all day animating hundreds of tentacles and trying to figure out how to prevent them from inter-penetrating,” notes Kramer. “We had to manage it and get through the shots. So one of our technical animators - Dan Sheerin - built a plug-in in Maya to build these tentacle limbs procedurally.”

Even maya lacks the ready to go tools to do it, had to create their own.

jeric_synergy
08-22-2017, 10:23 AM
An aside: I never quite 'got' why they were called 'Mimics'. --Not that I really care.

prometheus
08-22-2017, 10:47 AM
seen here, displacement wind


https://www.youtube.com/watch?v=bliPxn5e7mM

hmm...thought your answer to Mr Rid was a suggested alternativ to work similar as effectors, not sure that answer solves that, apart from displacement will most likely have tentacles intersect if used too much in strenght..right? and it isnīt controllable per tentacle bone or spline ctrl either.

I recall fiddling with making coralls, by simply making polylines, sending it to blender for skin meshing with skin modifier, send it back to lightwave and just apply displacement on the polylines and then metalink the mesh to that polyline, so you got wiggling underwater coralls, looks nice and is simple to do when you do not exceed the strenght too much so the tentacles do not intersect, and if you like to..you could instead of displacement on those polylines...just use bullet dynamics on the polyline chain, the metalink to the mesh will work and you can scale the collision size in bullet to be somewhat the same size at tentacles..roughly, and just add bullet wind effectors with their own vectors or use procedural textures in the forces.

The problem remains, making procedural motion while having it controllable with controllers at the same time as bullet is active, and if not bullet..you loose out on self collisions, I do not know if itīs possible to get spline control working with bullet, I do not think it is.

prometheus
08-22-2017, 11:52 AM
Double post...

prometheus
08-22-2017, 11:55 AM
just tested using bullet dynamics on polychains, branches made with line pen, then those are turned to bullet deforming bodies working on the polychain only, added ripple procedural displacement also on those branches, by default the bullet engine seem to override the displacement, so by changing map order to before world displacement ..you can have both procedurally animated tentacles as well as working with bullet to some degree, at least it works on static bodies, but should work on all the other forces and items as well in bullet, next step would be to deal with rigging a mesh with such polyline structure(at least tentacles) and then metalink that to those dedicated tentacles, and finally ..a way of controlling each line segment or some point/node ctrls for manual adjustment.

I havenīt investigated Ryan Royes tools and deformers for correcting dynamics after the simulation, then again chronosculpt just popped up as a question in my head..probably not feasable?

One way may be to go this way, after bullet sim on the polychain, scan it with clothfx, and enter edit mode, that way you may be able to move and correct some simulation, but itīs still very much limited.

erikals
08-22-2017, 12:45 PM
i was playing with the idea of something similar i think? (img)

moving away from the cloth/metalink idea though. seems better to mdd scan a rigged animation and add dynamics after.
or just add dynamics to main object.


The problem remains, making procedural motion while having it controllable with controllers at the same time as bullet is active, and if not bullet..you loose out on self collisions
yes, no easy way. tentacles in general > tricky.

erikals
08-22-2017, 01:10 PM
might be of use for other tentacle situations, plugin, SpringyThingy, realtime.
no self-collision, or dynamics though.


https://www.youtube.com/watch?v=BpFzPBx6Ewk

prometheus
08-22-2017, 01:24 PM
I miss ik booster with the dynamic bullet engine, I miss the bullet engine with clothfx options...

why o why couldnīt we have ik booster with proper dynamics, problem almost solved :)


https://www.youtube.com/watch?v=aAxA6XUU0c8&t=5s

prometheus
08-22-2017, 01:43 PM
erikals..by the way, something seems weird on youtube on my side, searching with lightwave as a search string, and filtering to upload date..your vids doesnīt show up anymore, I am almost certain yours should be shown.

erikals
08-22-2017, 02:49 PM
they look to be there, just on page 2

last idea is to mix 2 IKB simulations to make them collide.
might only work if baking the first tentackle set to mdd.

does IKB bone dynamics work on mdd objects?

erikals
08-22-2017, 02:56 PM
by the way > Davy Jones


https://www.youtube.com/watch?v=PhcrZyWNjlo

erikals
08-22-2017, 04:24 PM
just some thoughts... David Ridlen's idea in 2 takes.


https://www.youtube.com/watch?v=hpeYZzQ0f4E

hypersuperduper
08-23-2017, 10:10 AM
I had some luck with this solution: deform 2 point polychains with bones, and use a bullet deforming body on the 2 point poly chains with self collision on. Then just use metalink to tie quadrapus (?) mesh to the polychain mesh.

the setup makes it so that the bones tell the tentacles where they WANT to be but the physics dictates where they CAN be.

You get fast self collision that is pretty decent, it is reasonably stable and you can animate the bones however you see fit without worrying about intersecting tentacles . I just used ikbooster cause there is no setup. I am using a laptop, and I was able to manipulate the bones for 4 tentacles with dynamics on without it being completely unusable, and i could always just toggle dynamics off if it got too slow.

It isn't procedural per se, but the hard part is taken care of by dynamics, and with a better rig for the bones you could animate with even less effort.

https://www.youtube.com/watch?v=y1J-GqPnyuA

scene file:

137745

prometheus
08-23-2017, 12:59 PM
I had some luck with this solution: deform 2 point polychains with bones, and use a bullet deforming body on the 2 point poly chains with self collision on. Then just use metalink to tie quadrapus (?) mesh to the polychain mesh.

the setup makes it so that the bones tell the tentacles where they WANT to be but the physics dictates where they CAN be.

You get fast self collision that is pretty decent, it is reasonably stable and you can animate the bones however you see fit without worrying about intersecting tentacles . I just used ikbooster cause there is no setup. I am using a laptop, and I was able to manipulate the bones for 4 tentacles with dynamics on without it being completely unusable, and i could always just toggle dynamics off if it got too slow.

It isn't procedural per se, but the hard part is taken care of by dynamics, and with a better rig for the bones you could animate with even less effort.

https://www.youtube.com/watch?v=y1J-GqPnyuA

scene file:

137745


havenīt downloaded it..but that approach is what I actually had in mind as well..though I havenīt had the time to do it...will take a look at it, thanks.

erikals
11-03-2017, 03:45 PM
couldn't help but to think of this thread when i saw this animation  :)


https://vimeo.com/235440096

jwiede
11-10-2017, 08:43 PM
The big problem with using dynamics to drive tentacles is that they tend to lack "intention" (in animation terms). They don't seem to have much in the way of purpose behind the movements.

Watch videos of how a squid or octopus actually move their tentacles (and there are significant differences in how they do so, btw), the movements are typically highly coordinated and with clear purpose. Tentacles are extremely dextrous appendages, by definition. IMO, it'll be very difficult to get decent, rational-looking tentacle motion without explicit coordination and purpose -- even for non-squid/-octopus cases, such manipulators evolve to solve some explicit need, after all.

OTOH, if you're animating an inanimate "wacky wall-walker" or such, dynamics-based movement is probably fine.

MonroePoteet
11-11-2017, 03:53 PM
I'm a little reluctant to post the attached scene file, since the HardFX, ParticleFX and Animation Path wind used to implement the tentacles require careful execution for each render. But, it seems like at least something potentially useful in some cases, so here goes.

138508

138514 138509 138510 138515 138511
138512 138513

The basic implementation is to have a geometric Particle emitter using the Object-Vertices nozzle type, moving, rotating, twisting and scaling the emitter over time and then cutting off the emission when it grabs the bottle. In this case, I used a 7-sided flat disc, which is invisible to rendering (inactivated and non-visible in Scene Editor). In the sample scene, I sized the Emitter up fairly large to show the separate tentacles, but it could be kept scaled down to make a more dense "bundle" of tentacles, and twisted / sized in the direction of travel to make the tentacles appear to grow at different rates.

Surface-mode Hypervoxels are used to make the particles visible, the the HV size being a Gradient on Particle Age to make the tips of the tentacles sharp. Obviously, if you want blunter tentacles, the size at Particle Age 0 on the Gradient can be larger. I adjusted the "base" HV size to 50mm for my desired tentacle diameter.

An Animation Path type Wind was created by saving the final path of the Emitter as a Motion file, and then using MotionPathExtrude in Modeler to create a reference object, which is also inactive and hidden in Scene Editor after setting up the nodes. The Animation Path Wind was placed at the bottle's location, and then each successive node(n) Null (created by Animation Path and Cloned for each successive bend in the curve) to pull the tentacles back into the ground at their initial location. I didn't rotate and move the node(n) objects as the tentacles extrude and retract, but that'd be possible.

The Bottle is a HardFX object with Start at Collision set in the Collision tab. An invisible collision object hits the Bottle when the tentacles "grab" it, and then it follows the same Animation Path along the retracting tentacles into the ground.

Now, for the important parts:

To get the Particle density high enough to form tentacles, you can't use Calculate in the FX Tab of the Emitter's Property panel. You have to make a change to the Particle Emitter panel (I just re-select the Generate by Sec button to clear the internal cache), go to Frame 0, and hit the Play button. For whatever reason, this produces the high-density particles as requested by the Envelope on the Birth Rate, while Calculate doesn't. Huh.

And, to have the Bottle be affected by the Wind, you have to hit Calculate in the FX panel for it. So, it's IMPORTANT to do these things in this order when loading / changing the Scene:

1) Select the Bottle object
2) Go into it's HardFX tab, change at least one parameter (e.g. re-select Piece Mode as 1Piece)
3) Hit Calculate to calculate the Bottle's trajectory
4) Select the Emitter object
5) Go into it's ParticleFX tab, change at least one parameter (e.g. re-select Generate by Sec)
6) Go to Frame 0, and either scroll the timeline or press Play to generate the high-density particles

Maybe I'm missing something about the FX setup, but the above steps seem to work.

Good luck!
mTp

Surrealist.
11-11-2017, 08:14 PM
That is a very impressive effect, just by itself. Could you also then, animate that bottle, for instance, as if it was the foot of the character walking along the ground? And would there be a way to make that path follow something that resembled an IK chain? So effectively you could animate the character walking as that was happening. So the foot kind of grows out of the character and then walks along the ground. So I guess the tentacles would be actually reaching the bottle sooner. And the bottle and its path would move out from the character to the ground. Like in the mimic examples.

MonroePoteet
11-12-2017, 05:08 PM
That is a very impressive effect, just by itself. Could you also then, animate that bottle, for instance, as if it was the foot of the character walking along the ground? And would there be a way to make that path follow something that resembled an IK chain? So effectively you could animate the character walking as that was happening. So the foot kind of grows out of the character and then walks along the ground. So I guess the tentacles would be actually reaching the bottle sooner. And the bottle and its path would move out from the character to the ground. Like in the mimic examples.


Thanks! Well, I'd have to say "Yes and No". The setup required to do very much of it may be prohibitive, and getting the particle system to behave well each time you make a change to the FX objects is exacting and truthfully, a pain in the posterior. So, for example, the Scene and Object folder content I posted below was captured *IMMEDIATELY* after rendering the included video (honest!), but unless the steps (included below) are followed in sequence, you may not get the same results.

The included video looks MUCH better if slowed down by 1/2, 1/4 or 1/8th in VLC or UMPlayer:

138528

138529 138530
138531 138532

The attached sample scene uses the Genoma preset Rig Parts=>06_Legs=>Leg_Foot_01 with an exaggerated walk cycle. The Emitter was keyframed to migrate down the moving rig and back up, and then set to Post Behavior=>Repeat. It was stretched to 1.2x in all dimensions both top and bottom of the cycle to produce a sort of "hip" and "foot" for each cycle. I'd think it'd be feasible to automate the migration of the Emitter up and down the moving rig with an LScript, but I'd guess it'd be specific to the Genoma preset because of the rig-specific bone naming convention.

The Emitter had Heading and Pitch rotation keyframed with a Noisy Channel Modifier (Graph Editor=>Modifiers tab) to randomize the tentacle growth.

The Animation Path wind was created from the foot upward, parenting each subsequent node() (created by Animation Path or cloned) to each bone up the hierarchy. For the rig referenced above, the bone sequence (toe to hip) was:

dn(9)
dn(7)
dn(6)
dn(1)
dn(16)
dn(15)
dn(2)
dn(3)

for 8 handles total. The top and bottom handles were stretched to 1.5x or 1.2x to capture all created particles, or the tentacles start losing parts which float around stupidly.

All top-level objects were parented to the Master_Genoma_LegRig, so hypothetically the hierarchy could be cloned, but each cloned hierarchy would have to be walked to set targets, goals, etc.

To get the FX wind and particle system to behave consistently, you have to do the following in order each time you change either the Animation Path wind or the Particle emitter:




If changing the Animation Path wind, do the Calculate in the FX tab for it, then:
If changing the Particle emitter, get into the FX panel for the Emitter, change at least one parameter (e.g. re-click "Generate by" to Sec to clear the internal cache), then use either scrolling the timeline or Play to produce high-density particles (do NOT use Calculate on the FX tab). Note that Frame 0 may have "garbage" from the repeated emitter Birth Rate
Step forward a few frames
Press F9 to render - it seems to load some sort of VPR cache or something
Go into VPR mode - if what you see isn't what you expect, try repeating all steps (sigh!)


In the end, and I can't emphasize this enough, it may be easier to keyframe the motion of the Emitter for each tentacle "bundle" and produce the return Animation Path wind using the process I described in the previous post (save emitter motion, use MotionPathExtrude to produce reference object, align Animation Path nodes to reference object) rather than trying to use IK, but I could be wrong. It's the same technique, just different methods of determining the path of the Emitter and the Animation Path return path.

mTp

MonroePoteet
11-13-2017, 09:34 AM
Correction: In my previous post I said the nodes() of the Animation Path wind were *parented* to the bones going up the rig. They're not parented, they're set to IK in X,Y,Z and H,P,B and their *goal* is the bones up the rig. The nodes() have to be parented to the Animation Path wind object for them to calculate the wind effect correctly.

Also note that in order for the tentacles to return along their original path, the Animation Wind has to be shaped to the path using the method I described above (save emitter motion, MotionPathExtrude, etc.). It only works on the scene above because the Genoma rig is in the same plane and doesn't vary much from that.

mTp

Surrealist.
11-13-2017, 09:56 AM
Thanks for the additional info!

jeric_synergy
11-13-2017, 11:28 AM
Monroe, you da man.

MonroePoteet
11-14-2017, 10:30 AM
The attached sample scene uses the original method, but key-framing multiple 7-point emitters to create "child" tentacles with their Birth Rate offset until a little before they "sprout" from the base bundle.

For the Animation Path wind, instead of saving the emitter motion and creating a reference object for positioning the node() Nulls as described previously, I just set the Visibility of each emitter in turn to "Vertices" at the farthest extent of it's key-framed path, then use path of the emitted particles to place the Animation Path's nodes. Note that it's important to keep the Wind Power at 0.0% while positioning the nodes(), or the emitted particle path changes as the nodes are cloned and placed. If / when I get a chance, I'll experiment to see if a script to automatically create the "return" wind.

There are four Animation Path winds, one for each emitter, with the Group set to affect only that emitter's particle. As well, to make the particles vanish when they'd be entering back into the "body", there are four Collision objects set to Erase. Otherwise, the particles continue along the X-axis, which is not only unsightly but makes later frames very slow to render.

There's also a "spare" Emitter(5), which I had to add because I didn't think to parent the original bundle Emitter to a Null so it could keep rotating and expanding after stopping at its extent. Since the child emitters inherit the parent motion, it made the tentacles flail wildly. If I was creating the scene again, the FIRST step is to create a ChildTentacles null, parent the main bundle Emitter and all child Emitters to that.

Fine-tuning the maximum Birth Rate and HV characteristics for each emitter can take some time. This scene uses a 2000 particle/second birth rate, and Stretch to Velocity for the child tentacles with Blending turned on to form the tentacles.

Note that trying to use this technique with a parent "body" that's moving through space may (probably would be) problematic, since the HV's Stretch to Velocity capability uses World Coordinates rather than the Emitter's local coordinate system. So, if the tentacle-bundle's parent is moving through space, the HV's would stretch not only along the Animation Path wind's path, but would also stretch to the parent's motion and would probably look bad. I haven't tried it, but my experiment with the Genoma leg setup previously makes me think this is the case.

But, given all that, it looks pretty good to me, especially if Motion Blur was applied and it was part of an action shot with Camera motion, etc.

138554 138548
138549 138550
138551 138552

138553

mTp

inkpen3d
11-17-2017, 08:55 AM
This may have been mentioned before, but Jennifer Hachigian shows how she rigged a set of 8 tentacles using spline control and then animated them using just 4 master nulls in this SIGGRAPH 2013 video (at about 11 min 50 sec): https://www.youtube.com/watch?v=b_rFztpHYLc

Worth taking a look at.

Cheers,
Peter